INDEX

Explanations

names enclosed in quotation marks

oai_token-act-pair · gpt-3.5-turbo

quotation marks and attributed dialogue in the text

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 populated

-0.82

 permissions

-0.76

 setting

-0.76

 feminists

-0.75

 appointments

-0.75

 relate

-0.74

 consumers

-0.74

 defaults

-0.74

 undertaking

-0.73

 accessed

-0.73

POSITIVE LOGITS

Skip

1.34

Rocket

1.31

Wild

1.30

Doc

1.29

Fat

1.27

Big

1.23

Rust

1.22

Kid

1.21

Thunder

1.19

Bull

1.19

Activations Density 0.037%

names enclosed in quotation marks

quotation marks and attributed dialogue in the text

No Comments

No Known Activations